DR. SIMPSON is Executive Associate Dean for Academic Affairs, and Professor of Obstetrics and Gynecology and Human and Molecular Genetics, Florida International University College of Medicine, Miami, FL. He is Past President of the Society for Gynecologic

Joe Leigh Simpson, MD

Incorporating genetics into practice was once straightforward for our specialty. Prenatal genetic diagnosis was mostly relevant to women over age 35 at delivery, plus those in a particular ethnic group. Single genes seemed unlikely to be causative for more than exceptional cases. For example, searches for individual genes causing polycystic ovary syndrome (PCOS) and endometriosis floundered and proved hard to find for common cancers. We knew increased risks existed in first-degree relatives for some women with these conditions, but the etiologies were unclear and risks in these relatives were generally less than 5%. We invoked polygenic inheritance: many genes of small affect when present individually could cumulatively influence the presence or absence of the condition. In gynecological cancers, the applicable conditions involved breast, ovary and endometrium. However, management was, after all, the same whether the cause was familial or not. So, why bother delving into the arcane science of genetics?

Well, the genomic era and facile ability to sequence an individual’s cell- or tissue-derived DNA has upended our comfort zone. No longer can we remain aloof.

 in this issue by Leslie M. Randall MD, MAS, and Katherine Coakley, MD,  illustrates well the new reality. New conceptual approaches are here today for gynecological cancer and soon will be applicable for other conditions. With the cost of gene sequencing decreasing (perhaps one-third or less than an in vitro fertilization cycle) and the ease of bioinformatic interpretation increasing correspondingly, we can expect precision medicine to be inculcated in practices. We cannot hide any longer.

To appreciate our new responsibility, let’s take a step back and consider gynecological cancer and its heritability. We have long known that familial cases existed, but classic autosomal-dominant pedigrees were rare and the genetic mechanism was unknown. The relatively low recurrence risk in first-degree relatives dissuaded some of a single-gene etiology, consistent with reasoning that individuals with the same histology must have the same condition. Of course, we now know that all common, adult-onset disorders are heterogenous in their genetic etiology. Different genes and non-genetic etiologies contribute to the aggregate sum of cases. In gynecological cancer, the role of single mutant genes is now well-illustrated by hereditary breast and ovarian cancer (HBOC) and Lynch syndrome, an important cause of endometrial and colon cancer.

Clinically, some cases of gynecological cancer appear to be inherited, whereas others are sporadic. This was traditionally explained based on the assumption that cancer arises-whether inherited or not-by sequential mutations. The once touted “two-hit” hypothesis is less sacrosanct than before, but multiple hits are unequivocally involved. All this differs from genetic disorders like sickle cell anemia or cystic fibrosis, which require only a single “hit” in the requisite gene.

In heritable cancers in any organ, it is assumed that a parent has transmitted a mutant gene to offspring, even if he/she is asymptomatic. But there still must occur a second (or more) mutagenic “hit” for cancer to originate. One does not have to have a transmissible mutation to develop cancer, but if a person has one of the two or more required mutations, she has an unfortunate head start. In such cases, age at cancer onset is earlier than if two somatic mutations are required. Given that the transmitted mutation would have been in the embryo, the mutant gene would be present throughout the body (germline). Thus, it can be conveniently detected in the patient’s blood.

In contrast, the second hit must be specific to the organ (breast, ovary, endometrium) that will manifest cancer. Thus, one can detect this somatic mutation (not transmitted) only in the organ affected. This type of testing becomes applicable if genetic characteristics of the tumor dictate specific therapy.

In the genomic era, it became possible to sequence individuals and tissues, providing novel information concerning genes altered in a specific condition. Among the cohort of women with HBOC and Lynch syndrome it was possible to identify causative genes and determine the number of individuals having those genes. For example, 70% of patients with HBOC have 

 mutations. Faulty DNA repair genes-mismatch repair (MMR) genes 

-are causative in Lynch syndrome. Other mutations play roles, but less frequently.

Where do we stand now and what is different?

Simply put, the pedigree is now relevant. If a patient has a known transmissible mutation, her family members need to be tested. Family members, distant or close, qualify. If a clinically normal family member proves to have the (germline) mutation, a surveillance program should be initiated on the assumption that a subsequent “hit” will result in cancer. We have multiple laboratory options for testing for gynecological cancers. Should testing stop with common markers for the causative germline mutations (

; MMR)? Should multigene panels including less common mutations be ordered? Should we jump to whole-exome sequencing (all protein coding genes), which is actually less daunting and less expensive than one might imagine? I predict that whole-exome or genome sequencing will soon be preferable, not just for cancer testing but probably for all of us. Imagine the advantages of recognizing pharmacogenetic susceptibilities to avoid idiosyncratic drug reactions!

 In the 1990s, for good strategic reasons, the American College of Obstetricians and Gynecologists
established the goal of adding generalist medicine to our surgical background. Although a laudable goal, in my opinion, only a partial success. I suspect that our hearts were not in it. (After all, we went into ob/gyn to avoid having to deal ubiquitously with these conditions!) We have returned to this question in 2018, but now we really do bear the responsibilities.

Extending cancer testing to asymptomatic family members and following with surveillance truly is our domain. Actually, cancer testing is a good place to get into a comfort zone. We can predict that genetic differences will be correlated with different treatment protocols and different outcomes. We can anticipate that other gynecological conditions, such ovarian stimulation, endometriosis, and PCOS, will need to be approached in similar fashion.

For now, you should consider gynecologic cancer testing for HBOC and Lynch syndrome part of your day’s work.

Articles

Ob/gyns have a responsibility and can no longer be passive in this genomic era.

HBOC and Lynch syndrome testing: All in a day’s work

Once upon a time, medical advertising was considered quite inappropriate. In fact, physicians over a certain age may recall when directly communicating with the general public was limited to the Yellow Pages or perhaps understated announcements concerning retirements or new partners. This thinking now seems terribly antiquated, so last century. One is almost expected to advertise, and with great elan. Some may wistfully recall simpler times, but that was before we had the internet. 

And it's not just physicians who are advertising. Hospitals and pharmaceutical companies likewise peddle their wares to the public with great flare. Advertisements behind home plate remind us where sports injuries can be treated; time outs and injury updates during football games are sponsored by health systems. Clever television ads encourage backyard quarterbacks to benefit from sildenafil. It would be quixotic for us to resist this trend. The train has long since left the station. 

The latest direct-to-consumer (DTC) barrage, however, seems a bit more pernicious. The situation started with burgeoning genetic discoveries which, coupled with meteoric increases in technology, have permitted us to determine the risk of developing many disorders or suffering drug reactions. Testing for two variants in the cytochrome P-450 2C9 gene (CYP2C9) and one in the vitamin K epoxide complex 1 gene (VKORC1), for instance, can identify 50% of individuals who will experience hypersensitivity or resistance to warfarin.

 Similarly abacavir is a highly effective drug for managing human immunodeficiency virus (HIV) – infected patients, but 5% to 8% experience potentially fatal hypersensitivity. This can be accurately predicted on the basis of allele HLA-B*5701, the presence of which contraindicates abacavir therapy.

There seems to be a market for these DTC kits. One company offers a profile for prostate cancer, touting its ability to predict up to a 5-fold increased risk, or 9-fold if family history is taken into account. Another kit offers to assess one's nutritional needs. Even the sequence of the entire genome, or selected genes, can be ordered, a process approved by regulatory authorities in California.  

Given this onslaught of information, it is inevitable that at least some of our patients will be interested. Yet, they may not appreciate certain unintended consequences. An "adverse" genetic profile does not confer certainty of disease, nor the converse. What role then, should we obstetrician-gynecologists play in DTC testing?  

First, accept that patients will have DTC performed with or without our approval, and often inappropriately. Low risk patients are indeed influenced by media ads.

  Thus, we will be presented with test results post hoc, asked to explain their meaning, and recommend often unavailable preventative measures. 

Second, encourage patients to tell us of their interest in a given DTC test in advance, rather than merely waiting to present us with results. Does the patient truly understand that a DTC test may not provide a definitive answer, but only a relative risk? Or, that even a 20-fold increase in relative risk may confer a very low absolute risk? Or, that there are no treatment options at this time? Such discussion may result in ardor for a given test diminishing. Young women often ask for amniocentesis in order to have a "genetically perfect child". When they learn such a request is unrealistic, interest wanes.  

Third, the obstetrician-gynecologist needs to determine the extent to which he or she is comfortable in interpreting a given DTC result. If the disorder in question is outside our discipline, we could demur. Not so for DTC results concerning conditions like ovarian cancer. The American College of Medical Genetics offers useful minimal requirements for providers assessing direct-to-consumer (DTC) tests.

Patients are going to have direct-to-consumer genetic tests done, regardless of our position on the subject. At the very least, we can offer them sound advice on the matter.

DTC genetic testing: What do you tell patients?

Preimplantation genetic diagnosis (PGD) is a major advance, one that has given us the ability to diagnose a genetic disorder before implantation at 6 days postconception. For more than a decade, it's been an important addition to prenatal genetic diagnosis.

  Assisted reproductive technology (ART) is required, most commonly in vitro fertilization (IVF), followed by removing and analyzing a single blastomere at the eight-cell stage (3 days). Embryos subsequently classified as unaffected by genetic defects are transferred, and affected embryos are not. Analysis of polar bodies can usually accomplish the same goal. Because ART is required and the process is expensive and stressful, some indications are controversial, and guidelines are not always consistent or followed.

  My goal here is to explain the current indications for PGD, addressing these controversies. 

Avoiding clinical termination of affected pregnancies   

Of all the couples in the United States undergoing traditional prenatal genetic diagnosis, only half are prepared to terminate a pregnancy if their fetus is deemed abnormal. So on the one hand, many couples wish to be reassured that their fetus is normal, but on the other, they are not prepared to terminate an affected clinical pregnancy. Only PGD allows these couples the option of avoiding an abnormal child. PGD to avoid clinical termination has been a common attraction since the procedure was introduced in the 1990s; almost half of European couples undergoing PGD cite this as a reason.

  Even couples who have no religious objections to termination find traditional prenatal diagnosis daunting when termination becomes necessary in more than one pregnancy. In couples at risk for autosomal recessive disorders (e.g., cystic fibrosis), it's not rare to have consecutive fetuses affected (1/4 × 1/4 = 1/16). 

When pregnancy involves a prospective parent at risk for an adult-onset autosomal dominant disorder, PGD is the only pragmatic approach. This at-risk parent may be clinically normal but have a mutant gene not yet clinically manifested. Typically, the parent's at-risk status came to light because one of the prospective parents has a mother or father with such a disorder. If a patient's parent has the autosomal dominant disorder Huntington disease, for example, the risk for any given fetus is 1 in 4 (1 in 2 that the prospective parent is affected times 1 in 2 that, if so affected, the mutant gene will be transmitted). 

The at-risk parent may not wish to learn of his/her genotype, but may still want to avoid transmitting the mutant gene to the couple's offspring. Through PGD, the at-risk adult can do so, remaining unaware of his or her own status. The strategy requires screening all embryos, transferring only those unaffected, and never disclosing information on any individual embryo. The scenario must be repeated in subsequent cycles, even if tests reveal the at-risk parent to be unaffected. Otherwise, at-risk patients would readily deduce their genotype. 

Another unique capability of PGD involves identifying HLA-compatible embryos for transfer. Suppose a couple desiring a pregnancy has had a child with a lethal autosomal recessive disorder characterized by bone marrow failure. If stem cell transplantation could repopulate the child's bone marrow with genetically normal precursors, the affected sibling would be able to survive. Stem cell transplantation is highly successful if HLA-compatibility exists between donor and recipient, but considerably less successful when it does not. 

A sibling is the most likely source for HLA-compatible stem cells-including cells from a newborn sibling's umbilical cord blood. Although amniocentesis or chorionic villus sampling can identify HLA- compatible fetuses, only 1 of 4 fetuses 

 HLA compatible. The prospective fetuses will usually also be at risk for the autosomal recessive disorder causing the sibling's bone marrow failure; but 3 of 4 fetuses would be normal. What's more, the overall likelihood of a genetically normal HLA-compatible fetus is only 3 in 16 (i.e., a 1 in 4 likelihood of HLA-compatibility multiplied by a 3 in 4 likelihood of being unaffected). Only PGD provides a practical method for yielding an unaffected and HLA-compatible embryo. The most common indications are  β-thalassemia, Fanconi anemia, Wiskott-Alderich syndrome, and acute leukemia.

  Aneuploidy can be determined on the same embryos.

Among the clinical issues to consider: Should PGD aneuploidy testing be offered to women of advanced maternal age?

Preimplantation genetic diagnosis: Which patients to test and other controversies

But it is possible, because prenatal screening just isn't so perfect and some genetic conditions are relatively common. One in 800 liveborns has DS, and the incidence increases with advanced maternal age. One in every 2,500 Caucasians of Northern European or Ashkenazi Jewish ancestry has CF. Even if screening produces negative results, there 

 still reason for expectant parents to worry. The risk is far lower, to be sure, but affected cases do occur. 

How, exactly, do we find ourselves in the unenviable position of bearing bad news to a couple who believed their pregnancy was normal? The problem is our failure to help patients understand the difference between screening and case detection.  

In genuine genetic screening, by contrast, we acknowledge that not all cases will be detected. We accept detection rates of only 80%-and failure rates of 20%. But there are also certain prerequisites: a reliable assay, the possibility of altering management by curing a disease (neonatal screening) or permitting reproductive options, ability to address unexpected problems, and a favorable cost:benefit ratio.  

The standard of care for DS screening is analysis of selected serum analytes and measurement of nuchal translucency. Women whose risk is above a certain threshold-usually the 1 in 270 mid-trimester risk for a 35-year-old-are also offered chorionic villus sampling (CVS) or amniocentesis. Approximately 1 in 20 to 25 such invasive procedures will detect a fetus with DS.  

If we performed first-trimester CVS on just 5% of the women who have first-trimester screening, we could detect 87% to 93% of cases of DS. In the second trimester, using a triple screen would yield a 70% detection rate, and using a quadruple screen would yield a detection rate of 80%. The rate would increase to 95% with sequential screening in the first and second trimesters, if we could achieve full patient compliance. In fact, we could detect incrementally more cases of DS if we were willing to perform amniocentesis or CVS on more than 5% of pregnant women of all ages. That, however, would be unacceptable because of the cost:benefit ratio and the potential for complications. 

The only way to detect 100% of DS cases is to perform CVS or amniocentesis in every pregnancy. And we would then be practicing case detection, not screening. 

Many patients don't understand that genetic screening doesn't come with a guarantee. Yes, the consent forms they sign explain that truth. But emotionally, most mothers-to-be want to believe that detection is 100% accurate. All too often, we succumb to the desire to provide more good news than the data may warrant. Patients, too, are complicit in asking the same question in different ways: What are my results? Could the test be wrong? Have you ever seen a problem with this test? Rather than saying "Don't worry," perhaps a dose of caution would be more prudent.  

We need to remember that screening is screening; it's great, but not perfect.

Joe Leigh Simpson, MD

Articles

HBOC and Lynch syndrome testing: All in a day’s work

DTC genetic testing: What do you tell patients?

Preimplantation genetic diagnosis: Which patients to test and other controversies

Sign Out: Prenatal screening isn't perfect